Chroma subsampling

Chroma subsampling reduces the spatial resolution of color-difference data relative to brightness data in an image or video signal, exploiting human vision’s lower sensitivity to fine color detail.

YCbCr: A color representation that separates luma from two color-difference components. Subsampling is commonly applied to YCbCr’s color-difference components while retaining luma resolution.
Human visual system: The sensory and neural system that detects and interprets visual information. Its lower spatial sensitivity to color than luminance motivates subsampling.
H.264: A video-coding standard that compresses moving pictures for storage and transmission. H.264 commonly encodes video using 4:2:0 chroma subsampling.
4:4:4: A sampling format that stores a chroma sample for every luma sample, without chroma subsampling. It retains full chroma resolution, unlike subsampled formats such as 4:2:0.
Chroma: The colorfulness of a color relative to the brightness of a similarly illuminated white. Chroma is the color information reduced by subsampling.
Spatial frequency: The rate at which image intensity or color changes across space. Fine spatial detail is where reduced chroma resolution can become visible.
H.265: A video-coding standard designed to compress video more efficiently than H.264. H.265 streams often use subsampled chroma to reduce coded data.
4:1:1: A chroma-subsampling format that stores one chroma sample for every four horizontal luma samples. It reduces horizontal chroma resolution more aggressively than 4:2:2.
Luma: A weighted signal representing perceived brightness, commonly derived from red, green, and blue values. Subsampling preserves luma at higher spatial resolution than chroma.
Sampling theorem: A principle specifying the sampling rate needed to reconstruct a band-limited signal without aliasing. Subsampling lowers chroma sampling rates and therefore limits reconstructible color detail.

Linked from 12 pages